Correspondence 114 www.thelancet.com Vol 372 July 12, 2008 On the basis of all of the above points, one cannot conclude that probiotics in general present a risk in enteral feeding. I declare that I have no conflict of interest. Philippe Marteau philippe.marteau@lrb.aphp.fr AP-HP, Medico-surgical Department of Digestive Diseases, Lariboisière Hospital, 2 rue Ambroise Paré 75010 Paris, France 1 Besselink MGH, van Santvoort HC, Buskens E, et al, for the Dutch Acute Pancreatitis Study Group. Probiotic prophylaxis in predicted severe acute pancreatitis: a randomised, double-blind, placebo-controlled trial. Lancet 2008; 371: 651–59. 2 van Minnen LP, Timmerman HM, Lutgendorff F, et al. Modification of intestinal flora with multispecies probiotics reduces bacterial translocation and improves clinical course in a rat model of acute pancreatitis. Surgery 2007; 141: 470–80. 3 Marteau P, Shanahan F. Basic aspects and pharmacology of probiotics: an overview of pharmacokinetics, mechanisms of action and side-effects. Best Pract Res Clin Gastroenterol 2003; 17: 725–40. 4 Alfaleh K, Bassler D. Probiotics for prevention of necrotizing enterocolitis in preterm infants. Cochrane Database Syst Rev 2008; 1: CD005496. Authors’ reply Gregor Reid and colleagues correctly cite the definition of a probiotic provided by the Food and Agriculture Organiza- tion and WHO: “Live microorganisms which when administered in adequate amounts confer a health benefit on the host”. This does not mean that if such a product shows negative effects in a new study in a different population, the product should no longer qualify as a probiotic. From now on we will refer to the product used as a “combina- tion of probiotic strains” or “the combination” because the individual strains have shown beneficial effects in previous studies. Contrary to Reid and colleagues’ suggestion, our combination was designed after rigorous selection studies. 1 In several animal studies and smaller clinical studies, no negative effects of the combination were detected. As a consequence, we feel that it was probably not the combination but the administration of the combination together with the severity of the disease that was largely responsible for the effects obtained. The implication that the increased mortality rate in the probiotics group was caused by a higher rate of organ failure on the day of randomisation in the study product group is incorrect. None of the baseline characteristics differed significantly between the groups. Furthermore, in a post-hoc subgroup analysis, having excluded patients with organ failure on the day of randomisation, mortality was still twice as high and bowel ischaemia was significantly more frequent in the group receiving the study product. Bala Reddy and John MacFie propose “hypercaloric feeding” as a potential cause of bowel ischaemia. This was a post-hoc, quite heterogeneous, end- point and whether it was initiated by organ failure, inflammation, or other unknown factors is unclear. It is, how- ever, unlikely that it was related to hypercaloric feeding as such, because the feeding regimen—jejunal infusion of multifibre nutrition—was identical in both groups and the administration of the study product, with a negligible caloric load, was the only difference. Philippe Marteau raises concerns about the dose of probiotics, and we agree that studies on doses are in- deed lacking. However, several trials, in patients with acute pancreatitis and pa- tients scheduled for elective abdominal surgery, have used even higher doses of probiotics with a similar route of intrajejunal (bolus) administration. 2–4 As for our alleged conclusion that probiotics in general present a risk in enteral feeding, we feel that this does not accurately reflect our cautious inter- pretation that probiotics can no longer be considered harmless under all con- ditions, especially in critically ill patients. We declare that we have no conflict of interest. Marc G Besselink, Hjalmar C van Santvoort, Erik Buskens, Louis M Akkermans, *Hein G Gooszen, on behalf of the Dutch Acute Pancreatitis Study Group h.gooszen@umcutrecht.nl University Medical Center Utrecht, PO Box 85500, HP G04.228, 3508 GA Utrecht, Netherlands (MGB, HCvS, LMA, HGG); and University Medical Center Groningen, Groningen, Netherlands (EB) 1 Timmerman HM, Niers LE, Ridwan BU, et al. Design of a multispecies probiotic mixture to prevent infectious complications in critically ill patients. Clin Nutr 2007; 26: 450–59. 2 Rayes N, Seehofer D, Theruvath T, et al. Supply of pre- and probiotics reduces bacterial infection rates after liver transplantation— a randomized, double-blind trial. Am J Transplant 2005; 5: 125–30. 3 Rayes N, Seehofer D, Theruvath T, et al. Effect of enteral nutrition and synbiotics on bacterial infection rates after pylorus- preserving pancreatoduodenectomy: a randomized, double-blind trial. Ann Surg 2007; 246: 36–41. 4 Olah A, Belagyi T, Issekutz A, Gamal ME, Bengmark S. Randomized clinical trial of specific lactobacillus and fibre supplement to early enteral nutrition in patients with acute pancreatitis. Br J Surg 2002; 89: 1103–07. According to Marc Besselink and colleagues, the increase in mortality in patients with pancreatitis treated with probiotics was due to bowel ischaemia. 1 It has been suggested that most small-bowel diseases have a common final pathogenic pathway. 2 Accord- ingly, enterocytic damage by various assaults (biochemical, immunological, microbiologial, vascular, etc) leads to an increase in intestinal permeability. This increased permeability results in a tissue reaction as luminal sub- stances gain access to the mucosa where bacteria are the main neutro- phil chemoattractant. 2 The prototype of this damage is enteropathy caused by non-steroidal anti-inflammatory drugs, 3 but there are more than 30 situations in human beings in which an increase in intes- tinal permeability leads to a uniform- ly prevalent and severe enteropathy. 2 These enteropathies cannot easily be distinguished from each other, even by enteroscopy or histological- ly (on which the changes resemble ischaemia). The inflammation can respond to antimicrobials (metro- nidazole), implicating resident com- mensal anaerobic bacteria in the pathogenesis. 4 The patients described by Besselink and colleagues had pancreatitis, and